scriptableinstance updated per jat's comments. responses below.
http://gwt-code-reviews.appspot.com/1084801/diff/6001/7018 File plugins/npapi/ScriptableInstance.cpp (right): http://gwt-code-reviews.appspot.com/1084801/diff/6001/7018#newcode337 plugins/npapi/ScriptableInstance.cpp:337: char* out = (char*)NPN_MemAlloc(retStr.size() + 1); On 2010/11/08 20:00:00, jat wrote:
Instead use:
NPVariantProxy(*this, *result) = retStr;
or at least:
STDSTRING_TO_NPVARIANT(retStr, result);
NPVariantProxy was written to make sure that refcounting/and memory
ownership
was done properly. In this case, since it is only ever a string it
wouldn't be
too bad to just use the macro directly.
I believe what you are doing below would actually fail if there are
non-ASCII
characters in the string.
done. http://gwt-code-reviews.appspot.com/1084801/diff/6001/7022 File plugins/npapi/prebuilt/gwt-dev-plugin/background.html (right): http://gwt-code-reviews.appspot.com/1084801/diff/6001/7022#newcode34 plugins/npapi/prebuilt/gwt-dev-plugin/background.html:34: function devModeTabListener(tabId, changeInfo, tab) { On 2010/11/08 20:00:00, jat wrote:
Have you verified that the tab ID is constant across refreshes or
selecting a
bookmark in a given tab?
In my testing, it is constant for a given tab. Though, it doesn't really matter: the listener checks whenever a tab's URL is updated--regardless of the tabId. Note that this is purely informational: no permissions changes result from this pageAction. http://gwt-code-reviews.appspot.com/1084801/diff/6001/7029 File plugins/platform/Win/Preferences.cpp (left): http://gwt-code-reviews.appspot.com/1084801/diff/6001/7029#oldcode34 plugins/platform/Win/Preferences.cpp:34: static char* getAccessList(HKEY keyHandle) { On 2010/11/08 20:00:00, jat wrote:
Do we care about migrating existing user's preferences?
Ie, should we keep a variant of this code for a release, and when it
is first
run on Windows it loads all the preferences from the registry into the
Chrome
local storage?
I would vote no. I'd rather just move wholesale to the unified localstorage model. It's easy enough for the user to add them back in. Another conversation we may want to have is whether or not the rules will ever support actual patterns. As it is right now, it's an exact string match which makes the whole notion of exclusion a bit silly. http://gwt-code-reviews.appspot.com/1084801/show -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
